The specific aims for this collaboration are in two parts. The first deals with model development, and the second with using the model predictions. The aims are: 1. To develop more comprehensive system models using population kinetic software tools which will include known or potential covariates (resolving kinetic variability into fixed and random effects). 2. To use the model to: (a) estimate the means, covariances and confidence limits of the basic and derived model parameters, and (b) test for model validity through the use of specific statistical tests. Unexplored yet promising applications of population -based modeling exist in the fields of occupational and environmental toxicology. The database of toxicant disposition is fragmented amongst poisoning case reports, controlled exposure studies, industrial hygiene measurements of occupational exposures, and occupational and environmental epiderniologic investigations. No modeling paradigm currently exists for the integration of such data of varying completeness and quality. The use of physiologically-based kinetic (PBK) models to study toxicant disposition, to support risk assessment and to develop occupational exposure standards is currently limited by the lack of understanding of population variability. Recent attempts to describe model variability using Monte Carlo approaches have indeterminate applicability because they use presumed parameter distributions. Moreover, the current setting of safe exposure standards is constrained by subjective weighting of heterogeneous toxicologic data. In the area of drug disposition, application of the nonlinear mixed effects modeling to population phannacokinetics has been quite valuable in the characterization of clinical populations, which is useful for the design of safe dosing regimens. A similar approach using the software tools to be developed by RFPK to create a population-based PBK model would be a powerful tool to address these needs in toxicology.